High power Boost PFC inductor?

[treez]

How about winding three of these cores individually for 60uH and connecting then electrically in series ?
Its inefficient with regard to total wire length (and copper loss) but it would be an entirely practical solution that would take up hardly any extra board space.

thanks, for the idea, this would be great if there were off-the-shelf cradle mountings for the individual torroids, otherwise we would be stuck with having to custom mount all those torroids.

Thanks Mr Cool, I am looking up T300-52 ,

 

[FvM]

How about winding three of these cores individually for 60uH and connecting then electrically in series ?

The suggestion seems to overlook the difference between the series connection of three uncoupled versus three coupled inductors.

 

[treez]

The suggestion seems to overlook the difference between the series connection of three uncoupled versus three coupled inductors

thanks, I am trying to work out how you would couple the three torroids.

 

[Warpspeed]

Why do they need to be coupled ?
Three inductors in series produces three times the inductance.

 

[FvM]

Three inductors in series produces three times the inductance.

Yes. But a larger core produces the same inductance with less copper.

 

[treez]

T300-52

thanks for showing these Micrometals torroids, i must admit I tried to do it with one T30-66 and it saturates, then tried T30-M125 and it again saturates. I couldn't find T300-52 at first, but now have , so will try it.
I could try others from micrometals, but they would either be wider diameter than the kool mu, or need stacking just as i am doing anyway with the kool mu.

- - - Updated - - -

I calculate that we would have to glue 22 pieces of T30-66 torroids together to get 180uH inductance without saturating at the 39 Amps peak current

- - - Updated - - -

..regarding Micrometals, just tried there T300-40D, but it requires 147 turns to get 183uH...this will give too much winding loss. T300-52 is for above 50khz use.

Kool mu looks like the only one that can do it, -going to need some strong glue though, to hold two of them together.

 

[treez]

Hi Flux torroid core part number 58737A2 gives gives 180uH with 33 turns (single core), but with 18.3 watts of core loss.
The circuitousness of the iterative calculations means its a lot of "try and see".

 

[Warpspeed]

Yes. But a larger core produces the same inductance with less copper.

There are no larger Kool Mu cores, which is why he needs to stack three together.

As I said earlier in #19 :

Its inefficient with regard to total wire length (and copper loss) but it would be an entirely practical solution that would take up hardly any extra board space.

 

[FvM]

Didn't see the stock list, but according to the catalog, there are larger KoolMµ cores, both toroid and E shape.

 

[D.A.(Tony)Stewart]

Is this a result from consulting Mag-Inc's H/W support or using their software or trial and error?

 

[Velkarn]

I cant find a ferrite E core big enough for the 39 Amps peak current, and 180uH.

maybe use two or more inductors in series?

 

[szhighstar]

Below cores can be applied as PFC inductor:
1. Sendust
2. High Flux
3. MPP
4. Powder iron core
5. Mn-Zn
#1 and #2 and #3 and #4 are toroidal type normall, there is EE type for Sendust core.
#4 is EE or PQ or other type, it must be opened enough air gap for core, so that core is not saturation at Dc bias.
For these coress, price and loss and DC bias and Bm and curie are not same, it depends on your application, and select suitable core.

 

[treez]

Is this a result from consulting Mag-Inc's H/W support or using their software or trial and error?

We have calculated that the only way we an do it (using the smallest core) is to use two pieces of 77615A7 kool mu stacked, and then wind 39 turns round this, to give 180uH.
We did this by an iterative calculation involving Reluctance formula, and so forth. We had to convert to reluctance because that brings the area into it....the mag-inc website does not give calculation support for stacking of their cores.

 

[treez]

Hello,
This is a new question but slightly related to the previous posts here.
Its about any problems caused by inductors whose inductance changes as the current through them changes....
We have just designed a 450uH inductor for a 2000W boost PFC stage.
Ipk = 16.3 Amps
I(rms) = 10.6 Amps.
F(sw) = 45khz

Its 40 turns of litz wound ECW24 (7 strands) round two stacked kool mu torroids (part number = 77619A7)

B(pk) is just 269mT
Max B(pkpk) = 23mT

The core losses are just 1.4W

Since the kool mu material suffers variation in permeability with magnetising force, the inductance of this inductor actually changes over the 10ms mains half cycle from 896uH to 450uH.

Will this effect operation of the PFC stage at all. I can’t think it will but just wondering on the off chance.

 

[Easy peasy]

This effect is used a lot to gain smaller current ripple at lower powers, the gain of the pfc stage has to be adjusted to allow for the higher L, usually the res freq of Lmax & Co is not near the volt crossover freq ...

 

[mtwieg]

Will this effect operation of the PFC stage at all. I can’t think it will but just wondering on the off chance.

If you use average current mode control, it should not affect your PF so long as the bandwidth of the current control loop bandwidth is high enough (aim for 3kHz or higher). When designing the compensation for the current control loop, assume that the inductance is at its minimum, to be safe.

However if you were to use a peak current mode control with a transition mode controller, then you would see significant distortion in the input current, since those converters are based on the assumption that there is a specific ratio between peak and average current, because the inductance is linear. But for your power level, transition mode operation is not reasonable in the first place so this shouldn't be an issue.

 

[Mr.Cool]

ah, sorry my core suggestion was based on a design i did which was for 200kHz switching

 

[treez]

This relates to post #34
The FAN9672 PFC controller datasheet (page 6) says it can have a "Boost inductor mismatch" of +/-5%.
Do you know what this means?

FAN9672 datasheet:
https://www.fairchildsemi.com/datasheets/FA/FAN9672.pdf

- - - Updated - - -

ok, we've realised it means the mismatch between the two inductor values of the interleaved stage. Do you think a torroid powder inductor of the type in post#1 would be more likely to have mismatch than a gapped ferrite core type inductor?

 

[brushhead]

Hi,
I have just designed a 2kW PFC with a Kaschke 170.808. Works well. You may be able to use stacked cores as others have said.

Rob.

 

[treez]

thanks, the torroid of post#1 has good enough inductor matching so we will be ok with it, and those kool mu have superb thermal properties